Antepithecus Ameghino, 1901

Hitz, Ralph B., Flynn, John J. & Wyss, André R., 2006, New Basal Interatheriidae (Typotheria, Notoungulata, Mammalia) from the Paleogene of Central Chile, American Museum Novitates 3520 (1), pp. 1-32 : 13-24

Antepithecus Ameghino, 1901

Antepithecus brachystephanus Ameghino, 1901

Figures 11–12

REFERRED SPECIMEN: SGOPV 3604: partial cranium, with palate bearing right and left dP1-dP4, and right and left M1, and an erupting right M2, as well as portions of the right zygoma and rostrum.

LOCALITY AND DISTRIBUTION: Azufre Locality, Río Azufre valley, Abanico Formation, Chile. SGOPV 3604 represents part of a small suite of dental and postcranial material recovered in 1997 during helicopter access and reconnaissance prospecting of remote western exposures of the Abanico Formation along the Río Azufre, due north of the Río Tinguiririca drainage and west of Volcán Tinguiririca. Its taxonomic assignment indicates a Casamayoran age (likely ‘‘late’’ Barrancan subage, of Cifelli, 1985). Extremely steep cliff exposures along the west side of the north-south trending Río Azufre precluded access to most of the thick, nearly horizontal stratigraphic sequence exposed in the area during this brief exploratory trip. All of the specimens recovered from the area were collected in a large quebrada from float blocks of uncertain provenance.

BACKGROUND: We refer SGOPV 3604 from the Azufre Locality to Antepithecus ( Ameghino, 1901) . Antepithecus (1901) is a basal interatheriid from the Casamayoran SALMA; its dentition is nearly identical to Notopithecus . Simpson (1967) provided a thorough treatment and revision of Antepithecus , noting that its similarities with Notopithecus caused him to vacillate before ultimately concurring with Ameghino (1901) that Antepithecus indeed represents a distinct taxon. Simpson (1967) noted, in his diagnosis for Antepithecus , the following distinctions of Antepithecus (relative to Notopithecus ): ‘‘slightly more brachydont, protocone and hypocone less united, cheek teeth more elongate, horizontal mandibular ramus shallower’’ ( Simpson, 1967: 96). With respect to the greater elongation of the teeth in Antepithecus, Simpson stated elsewhere in his discussion (pp. 77–78) that the ‘‘p3, p4, m1, and m2 are significantly longer’’ and for the upper cheek teeth, ‘‘ A. brachystephanus averages longer than N. adapinus .’’ Simpson provided t-test support for this size distinction for some of the teeth in the lower dentition, but due to small sample size did not apply the statistical test to the upper dentitions.

Other than these highlights, Simpson provided a single detailed description of a dentition that served to characterize both Antepithecus and Notopithecus . Antepithecus is not yet known from cranial material. Following is a description of SGOPV 3604, and a discussion of its diagnostic resemblances to Antepithecus and differences from Notopithecus and other basal interatheriids. Mensural information is presented in table 3.

DESCRIPTION: The specimen consists of the right and left sides of a heavily worn dP1-dP4 series, recently erupted and unworn M1s, and a RM 2 in the process of erupting. Also exposed are the palate, a portion of the right zygoma, and a portion of the right rostrum (maxilla and premaxilla). Additional portions of the rostrum may be preserved within the matrix of the sample, but have yet to be prepared.

The primary criterion for identifying the dentition as consisting of dP1-dP4, and M1-2, is the relative wear on each tooth. The anterior tooth is less worn than the succeeding three, while the two posterior teeth show almost no wear (the last one is only partly erupted). This is the expected pattern of wear for a dP1-4, M1-2 series in notoungulates, where the eruption sequence for upper cheekteeth is dP2, dP3, dP4, dP1, M1, M2, M3, P2, P3, P4 (M. Bond, personal commun.). Evidently, this specimen is a juvenile in which the permanent premolars had yet to erupt.

The dP1 is worn, but still shows a prominent paracone and small parastyle. An ill-defined anterior emargination occurs low on the tooth. Wear has carved a basin between the paracone and protocone. Enamel is visible on all sides of the tooth but is completely absent on the worn occlusal surface.

The dP2 is a larger version of dP1, but with a definite low anterior cingulum and a more distinct parastyle.

The dP3 is likewise similar to the preceding two teeth, save for the presence of a small metacone.

The paracone and metacone of dP4 are of equal height (this tooth is less worn than the first three); an anterior cingulum occurs high (dorsally) on the crown, while the protocone and hypocone remain separate until the dorsal edge of the crown is reached (producing a lingual sulcus). A small cusp is present on the ectoloph between the paracone and metacone. The presence of this cusp defines a tiny basin on the ectoloph.

The nearly unworn M1 has a distinct paracone and parastyle with an inflection between these two structures labially. The paracone connects lingually with a less elevat- ed protocone. The tooth’s anterior margin bears a distinct cingulum low on the crown. The height of the metacone matches that of the paracone; strong undulation of the ectoloph reflects the large size of these two cusps and the inflection between them. The metacone connects lingually to the hypocone, the latter of which equals the protocone in height (the two lingual cusps being lower than the labial ones). The protocone and hypocone are deeply divided; likely they would have merged only after heavy wear. A posterior cingulum sits low on the tooth. Enamel covers the entire crown.

The only observable morphology on the erupting M2 is the emerging tip of the protocone, and a portion of the loph connecting the protocone to the paracone, the latter of which remains unerupted. Except for the complete lack of wear, the protocone and loph are identical to those seen on M1.

The palatine-maxilla suture is arcuate, reaching anteriorly to the middle of dP4. The maxilla is excavated anteriorly of the orbit, the infraorbital foramen being located above dP3. The zygoma bears a small descending process on its anteroventral border. In addition, it displays what we interpret as the suture between the maxilla and jugal. The latter element is reduced and excluded from the orbit, a hallmark interatheriid synapomorphy ( Riggs and Patterson, 1935).

DISCUSSION

With the modest amount of morphology available for study, especially with most of the preserved teeth being deciduous and with minimal occlusal wear on the only fully erupted permanent teeth (L & R M1), the precise taxonomic assignment of this specimen is tentative. SGOPV 3604 is clearly a basal interatheriid (see next paragraph) and in comparison to early interatheriids, a number of conditions distinguish this specimen from Notopithecus and support its referral to Antepithecus , including brachydonty, wellseparated hypocone and protocone, and size. In morphology and size SGOPV 3604 is extremely similar to Antepithecus brachystephanus , to which we assign the new material from the Azufre Locality, Chile. Simpson (1967) provisionally recognized a second species of Antepithecus , A. innexus , distinguishing it solely by its wider molars (a primitive feature, see below), but noted that it may represent an extreme variant of Antepithecus brachystephanus ; if A. innexus is a valid species, SGOPV 3604 differs from it in its narrower molars (and possibly the deciduous premolar features noted directly below). SGOPV 3604 does differ from specimens of Antepithecus brachystephanus , one of only two early interatheriid taxa preserving deciduous premolars (the other being Punapithecus minora ), in having a more pronounced protocone on dP1-4, resulting in teeth with longer lingual margins, and thus less triangular in outline, than in A. brachystephanus (although the proportions of the M1 are very similar to that of Antepithecus ). We decline to name the material of Antepithecus from the Azufre Locality as a new species at this time, because it would be diagnosed solely on differences in deciduous premolars; these are only rarely preserved, and their within-species variability is unknown. Recovery of additional specimens of the Chilean form (showing diagnostic differences in the permanent upper premolars or upper or lower teeth not represented in SGOPV 3604), or the recovery of comparable deciduous teeth of other taxa, may ultimately warrant recognition of a new species of Antepithecus , one most closely related to A. brachystephanus among known species of interatheriids.

Following the phylogenetic analyses of typothere relationships by Hitz (1995, 1997; see also Hitz et al., 2000), early diverging members of the minimally inclusive clade that includes Colbertia , Maxschlosseria, Brachystephanus , Tsamnichoria , Oldfieldthomasia , and Acropithecus are characterized by permanent upper molars that are significantly wider than long. In Oldfieldthomasia debilitata , for example, the upper molars are 30% wider than long. These dimensions impart an occlusal pattern that is likewise wider than long (following the very initial wear stage). In other typotheres (interatheriids, hegetotheriids, mesotheriids, and archaeohyracids) the paracone/parastyle region expands anteriorly, producing upper molars that are equidimensional (or even longer than wide). SGOPV 3604 has nearly equidimensional deciduous premolars and M1 (the only permanent cheektooth erupted in the specimen). SGOPV 3604 differs from mesotheriids, archaeohyracids, and hegetotheriids in being much smaller and very low crowned. The latter condition also distinguishes it from archaeopithecids. General dental morphology thus indicates that SGOPV 3604 represents a basal interatheriid (given its brachydont condition and lack of synapomorphies of later diverging interatheriid clades). The tentative recognition that the jugal is excluded from the orbit substantiates assignment to the Interatheriidae rather than any other typothere subgroup. The presence of an anterior cingulum on M1 of SGOPV 3604 is shared by only three interatheriid taxa: Notopithecus , Antepithecus , and Punapithecus . The extremely small size of Punapithecus argues against referring the new Chilean material to this taxon. Additionally, the M1 of Punapithecus has a parastyle that projects more anteriorly, giving the tooth a different aspect than SGOPV 3604, and the ectoloph of the M1 of Punapithecus is considerably smoother. Although the M 1 in SGOPV 3604 is nearly the same size as that of Notopithecus , the brachydont nature of this tooth (hypsodonty index or HI 5 0.45), plus the wellseparated protocone and hypocone, compare more favorably to Antepithecus . Recognizing the limited available diagnostic morphology, and the lack of recent taxonomic revisions or comparisons of Notopithecus and Antepithecus , the balance of evidence suggests that Antepithecus is the most reasonable assignment currently possible. AMNH 28701, Antepithecus brachystephanus , a juvenile specimen figured in Simpson (1967: 97) provides a useful comparison. The M1 of AMNH 28701 closely resembles that of SGOPV 3604, as do the deciduous premolars. The most notable distinction between AMNH 28701 and SGOPV 3604 is the more robust protocone on dP1-4 of the Chilean specimen, resulting in teeth with longer lingual (and hence less triangular) margins, than in A. brachystephanus . Although this difference might indicate a species-level distinction, lacking other diagnostic differences (particularly in the adult morphology or permanent dentition), and given the paucity of deciduous dentitions known for other basal interatheriids and poor understanding of within-species variability in deciduous premolar morphology, we choose not to recognize the Azufre Locality specimen as a new species.

SGOPV 3604 would thus seem to indicate a Casamayoran age (possibly the ‘‘late’’ Barrancan subage, Cifelli, 1985) for the strata in which it occurs along Río Azufre. Should the discovery of additional specimens (preserving M3, permanent upper premolars or other anterior teeth, or lower dentitions) tilt the taxonomic assignment of the Azufre material toward Notopithecus (unlikely, given their many morphological differences) rather than Antepithecus , the biochronologic implications remain the same, as both Antepithecus and Notopithecus occur throughout approximately the same temporal interval. With a few (likely erroneous) exceptions, both Notopithecus and Antepithecus are restricted to the Casamayoran SALMA. Simpson (1967) listed Notopithecus as occurring within the Riochican, but Cifelli (1985) questioned this assigment, identifying the suspect specimens as Notopithecinae indet. Although Simpson (1967) and Marshall et al. (1983) both listed? Notopithecus as occurring in the Mustersan SALMA, these authors obviously were not convinced of this allocation. Thus, the range of specimens securely identifiable as Notopithecus and Antepithecus is restricted to the Casamayoran. Moreover, both taxa are limited to the Barrancan subage of the Casamayoran, the younger of the two subages identified by Cifelli (1985) on the basis of exposures at two classic localities, the Gran Barranca and Cañadón Vaca. Cifelli (1985) further divided the Barrancan subage into ‘‘early’’ and ‘‘late’’ intervals. With one exception (see below), Notopithecus appears only in the ‘‘early’’ Barrancan and Antepithecus in the ‘‘late’’ Barrancan. In the uppermost portion of Section 2 from Cañadón Vaca, Cifelli (1985) reported the isolated occurrence of Notopithecus , indicating that the range of this taxon extends very slightly into the preceding Vacan subage. Kay et al. (1999) used new radioisotopic data to correlate the Barrancan subage of the Casamayoran SALMA to two possible paleomagnetic chrons, giving an age range of 35.34–37.60 Ma for this section, and an entirely late Eocene age for the Casamayoran (as opposed to its traditional conception as early Eocene, see Flynn and Swisher, 1995). Kay et al. (1999) did not directly address the age of the Vacan in the Cañadón Vaca section.

INTERATHERIID PHYLOGENY AND THE RELATIONSHIPS OF THE CHILEAN BASAL INTERATHERIIDS

Several cladistic studies of the Interatheriidae have been completed in recent years ( Cifelli, 1993; Hitz, 1997; Reguero, 1999; Reguero et al., 1996, 2003b), with varying levels of success with respect to resolution of interathere phylogenetic relationships. Cifelli’s work, based almost exclusively on dental characters, identified epidemic homoplasy in the group; the single resolved node stemming from his study was one demarcating the Interatheriinae from Notopithecus . Cifelli (1993) suggested, as in previous papers ( Cifelli, 1985; MacFadden et al., 1985, Marshall et al., 1986), the probable existence of two monophyletic groups within the Interatheriinae : ‘‘ Interatherium and allies’’ and ‘‘ Protypotherium and allies’’. Cifelli also highlighted four provisional synapomorphies for Interatheriidae : maxilla excluding jugal from orbit (sensu Riggs and Patterson, 1935), I 1 -C bifid, I 2 -P 1 transversely compressed, lower molars bilobate.

Hitz (1997; the phylogeny stemming therefrom is presented in Hitz et al., 2000) proposed a more highly resolved interatheriid phylogeny based on an analysis of 44 dental and cranial characters across 17 taxa. This more comprehensive level of sampling, and the degree of resolution achieved, motivated a phylogenetic definition (node-based) for the name Interatheriinae to accompany the list of synapomorphies diagnostic of the clade to which this name was associated ( Hitz et al., 2000).

Reguero et al. (2003b) presented a phylogenetic analysis of Interatheriidae , focusing mainly on interatheriine taxa ( Notopithecus representing the only basal interatheriid), building on the studies of Reguero et al. (1996) and Reguero (1999). Reguero et al. (2003b) achieved well-resolved relationships within Interatheriinae , although differing somewhat from those proposed by Hitz (1997), Hitz et al. (2000), and herein. Possible sources of these discrepancies are discussed below.

Here we present an analysis of interatheriids based on the data set used in Hitz (1997) augmented with the information available for

TABLE 4

Taxa Studied for Phylogenetic Analysis

Archaeophylus patrius Ameghino, 1897

Plagiarthrus clivus (5 Argyrohyrax proavus ) Ameghino, 1897

Progaleopithecus tourneri (5 Argyrohyrax proavunculus ) Ameghino, 1904

Protypotherium australe Ameghino, 1887 (see also Sinclair, 1909)

Protypotherium praerutilum Ameghino, 1887 (see also Sinclair, 1909)

Interatherium robustum Ameghino, 1887 (see also Sinclair, 1909)

Interatherium extensum Ameghino, 1887 (see also Sinclair, 1909)

Cochilius volvens Ameghino, 1902

Cochilius fumensis Simpson, 1932

Miocochilius anamopodus Stirton, 1953

Punapithecus minor Reguero et al., 1996

Santiagorothia chiliensis Hitz et al., 2000

Proargyrohyrax curanderensis Hitz et al., 2000

Notopithecus adapinus Ameghino, 1897

Transpithecus obentus Ameghino, 1901

Guilielmoscottia plicifera Ameghino, 1901

Colbertia magellanica Price and Paula Couto, 1950

Eopachyrucos plicifera ( Ameghino, 1901)

Eopachyrucos ranchoverdensis ( Reguero et al., 2003b)

Salla new taxon A (informal taxon, Hitz, 1997)

Salla new taxon B (informal taxon, Hitz, 1997)

five additional taxa: 1) Punapithecus , a basal interatheriid from northwestern Argentina ( López and Bond, 1995); 2) Proargyrohyrax , a recently described interatheriine ( Hitz et al., 2000); 3) Antepithecus brachystephanus ; 4) SGOPV 3604 (treated in isolation from Antepithecus and Notopithecus ); and 5) Eopachyrucos (the diagnosis of which has been recently emended [ Hitz et al., 2000], including a recently described species E. ranchoverdensis [ Reguero et al., 2003b]) (see table 4 for a list of taxa in the analysis). Colbertia , an oldfieldthomasiid, served as the outgroup. The character list and data matrix (appendix 1) included 43 dental and cranial characters across 21 taxa. The analysis, using PAUP 4.0 ( Swofford, 2002), yielded 623 equally parsimonious trees of 70 steps (branch and bound search, all characters unordered, equally weighted; for individual trees: CI 5 0.74, RI 5 0.86). All characters discussed below are numbered as in the character matrix.

Several nomenclatural issues regarding the names Plagiarthrus Ameghino, 1896 , Argyrohyrax Ameghino, 1897 , and Progaleopithecus Ameghino, 1904 , require clarification. All recent sources agree that the name Argyrohyrax is a junior synonym of Plagiarthrus ( Patterson, 1952; Marshall et al., 1986; Reguero, 1999; Reguero et al., 2003b), as suggested initially by Loomis (1914) and Simpson (1932). This synonymy was documented most concretely by Patterson’s (1952) observation that Argyrohyrax proavus (the type species of Argyrohyrax ) is based on an upper dentition belonging to Plagiarthrus clivus (the type species of Plagiarthrus ). Following Hitz (1997) and Hitz et al. (2000), we use the name Plagiarthrus to refer to specimens of Plagiarthrus clivus as well as those previously referred to Argyrohyrax proavus .

Secondly, Patterson (1952; repeated by Marshall et al., 1986) suggested that the type specimen of Progaleopithecus tourneri Ameghino, 1904 , probably represents a portion of the same specimen upon which Ameghino based Argyrohyrax proavunculus , rendering Progaleopithecus a junior synonym of Argyrohyrax . This synonymy was accepted by Hitz (1997) and Hitz et al. (2000), wherein species previously placed in Progaleopithecus (sensu Ameghino) were transferred to Argyrohyrax .

Lastly, Reguero (1999) and Reguero et al. (2003b) further revised interatheriid nomenclature, although some of their recommendations are at odds with the points just discussed. While recognizing the general synonymy of Argyrohyrax and Plagiarthrus , these authors regard Plagiarthrus as having priority (over Argyrohyrax ), and recognize proavunculus as a distinct species within Plagiarthrus — proposing the new combination Plagiarthrus proavunculus Reguero 1999 . Although unable to examine specimens of Argyrohyrax proavunculus directly, we note that in Ameghino’s original descriptions (1897) of Argyrohyrax proavus and Argyrohyrax proavunculus he stated that A. proavunculus is merely a smaller version of A. proavus and otherwise is very similar in morphology. Accepting A. proavus as a junior synonym of Plagiarthrus clivus justifies the new combination Plagiarthrus proavunculus ( Reguero, 1999) for ‘‘ A. proavunculus ’’. Reguero (1999) and Reguero et al. (2003b) also suggested, based on the presence of diagnostic characters (bicolumnar lower incisors, rounded [?molar] trigonids; Reguero et al., 2003b), that Progaleopithecus should continue to be recognized as distinct, rather than as a synonym of Argyrohyrax proavunculus as suggested previously by Patterson (1952) and Marshall et al. (1986). Having examined specimens of Progaleopithecus tourneri firsthand, we concur that there are sufficient diagnostic characters to justify continued recognition of this taxon. Here we use the name Plagiarthrus to refer to Plagiarthrus clivus (5 Argyrohyrax proavus ) and Plagiathrus proavunculus , while Progaleopithecus refers to P. tourneri , which is distinct from Plagiathrus proavunculus (5 Argyrohyrax proavunculus ).

The differences between the maximally parsimonious hypothesis of relationships identified here and the results from our previous analyses ( Hitz, 1997; Hitz et al., 2000) primarily center on decreased resolution of the basal part of the entire tree. Three of the taxa added to the current analysis are very fragmentary with respect to the character matrix; for example, SGOPV 3604 is only 13% complete, Punapithecus is 46% complete, and Proargyrohyrax is 50% complete. None preserves a significant amount of cranial material. Novacek (1989, 1992a, 1992b) documented that missing data contributed to tree instability insofar as greater numbers of maximally parsimonious trees resulted from the inclusion of taxa with considerable missing data. The missing data associated with SGOPV 3604, Punapithecus , and Proargyrohyrax seem to account for the decreased resolution from our earlier results. We have opted to employ the complete character matrix rather than exclude fragmentary taxa that potentially could act as ‘‘wildcards’’ ( Nixon and Wheeler, 1992) or otherwise obscure phylogenetic relationships, for several reasons. First, the correlation between missing data and ‘‘wildcard behavior’’ is not well demonstrated (Kearney and Clark, 2003). Second, the results of our analysis, although more polytomous than in the analysis by Hitz (1997) still identify many well-supported clades. Moreover, as it happens, some of the taxa we are most interested in placing phylogenetically (e.g., SGOPV 3604) are also among the most fragmentary.

An important contrast between the results of the present analysis and those presented by Hitz (1997) is that Santiagorothia chiliensis (as a consequence of adding new taxa to the analysis) is no longer resolved as the sole proximal outgroup to all other previously recognized interatheriines. This raises some questions with respect to the phylogenetic definition for the name Interatheriinae (sensu Hitz et al., 2000) considered below.

A strict consensus tree (fig. 6) identifies a clade (node 1 in fig. 6) that we equate with Interatheriidae . Monophyly of this clade is supported by many unambiguous synapomorphies: I 2 -C laterally compressed (sensu Cifelli, 1993) (character 2: character state 1), jugal excluded from orbit by maxilla (sensu Cifelli, 1993; Riggs and Patterson, 1935) (character 30: character state 1), nasofrontal suture anterior of orbit margin (character 31: character state 1), descending process on zygoma (character 33: character state 1 or 2), zygoma angled steeply at juncture with face (character 36: character state 1 or 2), narrow muzzle (character 37: character state 1), glenoid fossa deeply excavated and narrow (character 41: character state 1), postglenoid process sharp and narrow and separating fossa from meatus (character 42: character state 1 or 2). Two additional features identified by Cifelli (1993) (I 1 -C bifid, lower molars bilobate) as characterizing Interatheriidae are less widely distributed and therefore are not regarded as synapomorphic for all interatheriids in this analysis.

Numerous clades within Interatheriidae are supported by unambiguous synapomorphies. (There exist additional, potential synapomorphies diagnosing the nodes in fig. 6—beyond those listed. Nevertheless, these are considered equivocal, as at least one basal taxon of the clade that the character potentially diagnoses, and the nearest outgroup of that clade, cannot be scored.) Of note is the clade (the polytomy of node 4 in fig. 6) that is roughly equivalent to—aside from including Proargyrohyrax — the Interatheriinae (sensu Hitz et al., 2000, p. 3, defined as: ‘‘the clade stemming from the most recent ancestor of Santiagorothia chiliensis and Interatherium plus all of its descendants’’). Following the phylogenetic definition of Hitz et al. (2000), Proargyrohyrax is thus presently considered a member of the Interatheriinae . Should future work resolve the polytomy and place Proargyrohyrax as an outgroup to the least inclusive clade containing Santiagorothia chiliensis and Interatherium , then Proargyrohyrax would no longer be an interatheriine, by definition ( Hitz et al., 2000). Thus, new results might yield changes in the phylogenetic placement of taxa ( Proargyrohyrax in this case), without perturbing in the least a stable definition of the widely used name Interatheriinae . Three subgroups within the above clade warrant discussion. The first (node 5 in fig. 6), including Interatherium , Cochilius , and Plagiarthrus , is diagnosed by talonids on p2–4 subequal or larger than the trigonids (character 23: character state 3). The second (node 6 in fig. 6) is an unresolved tritomy of Protypotherium , Progaleopithecus tournoueri , and Miocochilius , diagnosed by a distinct third lobe on m3 (character 28: character state 1) and a narrow anterior border on the lower premolars and molars (character 29: character state 1). These two subgroups appear to correspond to the ‘‘ Interatherium and allies’’ and the ‘‘ Protypotherium and allies’’ of Cifelli (1993). A clade (node 2 in fig. 6) relevant to the systematic discussions in this paper is the least inclusive one containing Interatheriinae (sensu Hitz et al., 2000) plus Eopachyrucos , Johnbell hatcheri , and Ignigena minisculus ; it is diagnosed by two synapomorphies: M1–3 paracone/parastyle inflection is reduced or absent (character 12: character state 1), and upper molars longer than wide (character 17: character state 1).

In a second analysis, all trees # 71 steps (i.e., one step more than the maximally parsimonious) were retained, and a strict consensus tree derived from them. This process yielded 12,548 equally parsimonious trees. Several clades of note persist in this consensus tree: the pairing of Interatherium and Cochilius , a clade consisting of Interatheriinae (sensu Hitz et al., 2000) plus Proargyrohyrax and Eopachyrucos (node 3, see fig. 13), and a clade consisting of Johnbell hatcheri and Ignigena minisculus plus the previous clade (node 2, see fig. 13). The former clade (node 3) is the only one to persist in a third analysis, in which all trees # 72 steps were retained (126,998).

We next compare our results with those of Reguero et al. (2003b) to the extent possible given differences between the datasets (fig. 14). Reguero et al. (2003b) used 29 characters (22 dental; seven mandibular, cranial, and postcranial) across 12 taxa. Ten of the taxa in their analysis were interatheriines, and one was a basal interatheriid. They presented a single consensus of their most parsimonious trees, but they did not run additional analyses retaining trees of greater length. By contrast, we employed 43 characters (29 dental; 14 mandibular, cranial, and postcranial) across 21 taxa (11 interatheriines, 9 basal interatheriids, and 1 outgroup). Reviewing the characters in Reguero et al. (2003b), we determined 18 of them as roughly coincident with characters in our dataset, leaving 11 characters not corresponding to any of ours. Conversely, 25 characters in our dataset have no equivalents in Reguero et al. (2003b). The main difference between the consensus trees presented in Reguero et al. (2003b) and this study is the topology of the clades within the Interatheriinae (sensu Hitz et al., 2000) . Reguero et al. (2003b) placed Plagiarthrus and Progaleopithecus in more basal positions within Interatheriinae than did our results.

Both studies identify Eopachyrucos as having diverged prior to the appearance of the interatheriine (sensu Hitz et al., 2000) common ancestor. Reguero et al. (2003b) identified Proargyrohyrax as the nearest outgroup to Interatheriinae (sensu Hitz et al., 2000) , whereas in the current analysis Proargyrohyrax falls in a basal interatheriine tritomy. Although Eopachyrucos was considered an interatheriine by both Hitz et al. (2000) and Reguero et al. (2003b), based on its very high-crowned cheekteeth, it cannot be considered such any more, given that it now lies as an outgroup to the clade linked to the name Interatheriinae, sensu Hitz et al. (2000) . The shifting taxonomy of Eopachyrucos presents an opportunity to discuss the usefulness of applying phylogenetic definitions to groups the understanding of whose phylogenetic relationships is in flux, and why a phylogenetic definition is of particular utility in the case of the name Interatheriinae .

A potential criticism of the phylogenetic definition for Interatheriinae (sensu Hitz et al., 2000) is that it is now at odds with the distribution of one of the characters commonly used to diagnose interatheriines traditionally, high-crowned cheekteeth ( Cifelli, 1985). We explore an alternative to our earlier phylogenetic definition ( Hitz et al., 2000), one possibly better conforming to classical usage, concluding, however, that our original node-based definition is as robust as any conceivable alternative and that its continued use enhances the stability of interathere taxonomy.

An alternative to a node-based phylogenetic definition (as that of Hitz et al., 2000) would be an apomorphy-based definition ( de Queiroz and Gauthier, 1990, 1992), with hypsodonty being the logical candidate as the specifying apomorphy. A hypothetical apomorphy-based phylogenetic definition might read, ‘‘the clade stemming from the most recent common ancestor of all interatheres marked by hypsodonty’’. The name would refer to a clade encompassing all interathere taxa historically considered to be interatheriines (those that are hypsodont), plus it would be in alignment with moderately robust nodes in both the phylogenetic analyses of Reguero et al. (2003b) and this study (node 3 in fig. 6). [An extension of this example would be to use any of the diagnostic characters associated with the clade named Interatheriinae .]

We perceive two weaknesses with this alternative, the first being that hypsodonty is ultimately a gradational designation. Hypsodonty indices vary continuously from very low to very high; it does not occur in discrete states.

The second, more fundamental, problem is that with the broad sampling of interatheres in our analysis, we observe that the morphological contrast between what have traditionally been termed interatheriines and their nearest outgroups has been lessened, with the continuing recognition of successive outgroups forming a ‘‘comb’’ (e.g., nodes 2, 3, and 4 in fig. 6). Morphological ‘‘jumps’’ along this comb are small (i.e., nodes with significantly more supporting character changes than preceding and following nodes). Indeed, the characters identified by Hitz et al. (2000) as diagnostic of Interatheriinae are now spread across nodes 2, 3, and 4 (fig. 6). Santiagorothia chiliensis , Proargyrohyrax , and Eopachyrucos , all of which have been recently described (or emended), may be construed as morphologically intermediate between basal interatheriids and classic interatheriines such as Interatherium and Protypotherium . This is well illustrated by the consensus of trees 71 steps or shorter (fig. 13), which identifies the former three taxa as sharing a unique common ancestry with (but being placed outside of) the clade containing all other interatheriines. In this case, new discoveries clearly document that traits once thought to arise coincident with the origin of ‘‘interatheriines’’ (possibly as a correlated suite of adaptations to grazing or rapid environmental change) actually were acquired sequentially or in a mosaic fashion across a longer time span, and only appeared clustered because of substantial taxonomic and temporal sampling gaps. In fact, recovery of morphologically intermediate taxa is precisely the expected outcome of sampling new temporal intervals or previously poorly understood portions of the phylogeny (recent discoveries of a long series of transitional feathered dinosaurs are an appropriate analogy).

One could argue that the name Interatheriinae has been used historically to refer to taxa clumping at one end of a morphological spectrum ranging from early, small, brachydont basal interatheriids to later, larger, hypsodont or hypselodont forms, but without a clear demarcation of the taxonomic ‘‘dividing point’’. The increasingly blurred ‘‘boundary’’ between traditional ‘‘interatheriines’’ and their proximal outgroups highlights the many clades available to chose from in crafting a phylogenetically defined version of that name.

In the interest of stability, and in the absence of a stronger alternative, here we opt to leave the phylogenetic definition of Interatheriine (sensu Hitz et al., 2000) unchanged. To modify the definition to better conform with the traditional delineation between basal interatheres and interatheriines is to chase an historical ideal that, in light of recently described ‘‘intermediate’’ taxa, no longer exists.

There are several consequences of this action. First, the character diagnosis of the node associated with the name Interatheriinae in Hitz et al. (2000) changes to reflect the present analysis (distinctly bilobed p3-4 with persistent labial and lingual sulci; maxilla excluded from the superior orbital border by an anteriorly projecting sliver of frontal; auditory bulla lapped posteriorly onto the paraoccipital process—the latter two characters are provisional, as proximal outgroups to the clade in question cannot be scored for these characters). Second, Eopachyrucos , previously considered an interatheriine ( Hitz et al., 2000; see also Reguero et al., 2003b) based on the phylogeny and diagnosis of the clade accepted at that time, is currently seen as falling outside the clade to which that name is tied. Our conception of the clade’s membership has thus changed slightly since the phylogenetic definition was first proposed.

Our current phylogenetic results, and those of other studies (e.g., Cifelli, 1993; Reguero et al., 1996, 2003b; Hitz, 1997), underscore the strong support for the monophyly of Interatheriidae , regardless of the precise list of synapomorphies diagnosing it. In these phylogenies, Interatheriidae is consistently diagnosed by the apomorphic sandwiching of the jugal between the maxilla and squamosal, plus the exclusion of the jugal from the orbit, unusual features not seen in other notoungulates and that have long been used to assign taxa to the Interatheriidae . Other conditionally synapomorphic attributes are listed above and in Cifelli (1993). As the name Interatheriidae has not been defined phylogenetically, we do so here.

The unique configuration of the zygoma in interatheres offers a convenient basis for defining the name of this clade of notoungulates via an apomorphy-based phylogenetic definition. Defining the name Interatheriidae in this manner maximizes its congruence with historical usage. In apomorphy-based definitions the appearance of a particular character provides the means of stipulating the specific clade/ancestor to which a given name is linked. Interatheriidae is hereby defined as the notoungulate clade stemming from the ancestor displaying a splintlike jugal on the zygomatic arch, wherein this element is ‘‘sandwiched’’ between the maxilla and squamosal and excluded from the orbit.

All cladistic analyses of the Interatheriidae to date (e.g., Hitz, 1997; Reguero et al., 2003b; this study) highlight the paraphyletic nature of the Notopithecinae . Ameghino (1897) coined the name Notopithecidae to encompass solely Notopithecus adapinus , a taxon he considered ancestral to ‘‘prosimians’’ and other primates. More credibly, within the same paper he also recognized the close relationship of Notopithecidae to the Interatheriidae (5 ‘‘Protypotheriidae’’). Although the proposed link of Notopithecus to Primates has long been outmoded, Riggs and Patterson (1935) persuasively confirmed the affiliation of Notopithecus and interatheriids by virtue of the unique configuration of the jugal mentioned previously. Attempting to highlight this anatomical finding, Simpson (1945) proposed the subfamily Notopithecinae (within Interatheriidae ) to encompass Notopithecus , Antepithecus , Transpithecus , and Guilielmoscottia — to emphasize their lack of derived features such as hypsodont teeth, a taxonomic grouping that remains in wide usage to this day (e.g., McKenna and Bell, 1997).

Simpson (1967) provided a comprehensive systematic treatment of Notopithecinae . His ‘‘definition’’ (diagnosis—in current terminology) of the group listed a mix of cranial and dental characters, some being primitive features that also are typical of other typotheres (such as oldfieldthomasiids and archaeopithecids), while others are more likely synapomorphic for Interatheriidae . No synapomorphies uniting ‘‘notopithecines’’ to the exclusion of other interatheriids have yet been identified. We propose that non-interatheriine interatheriids are more appropriately termed ‘‘basal interatheriids’’ as this phrasing lacks any connotation of the monophyly of those taxa. The name Notopithecinae (and all of its derivatives) may someday be useful if any interatheriid species prove to be more closely related to Notopithecus than to other interatheriids. In such a case, this vintage name might be linked to this newly recognized clade.

Figures

Fig. 11. Antepithecus brachystephanus, SGOPV 3604 (referred specimen), palate, right and left dP1-dP4, right and left M1, erupting right M2, portions of the right zygoma and rostrum, occlusal view. Azufre Locality.

Fig. 12. Antepithecus brachystephanus, SGOPV 3604 (referred specimen), right dP1-dP4, right M1, erupting right M2, portions of the right zygoma and rostrum, right buccal view. Azufre Locality.